Spawn JupyterHub single user notebook servers using Ansible.
This spawner runs Ansible playbooks to start, manage and stop JupyterHub singleuser servers. This means any Ansible module can be used to orchestrate your singleuser servers, including Docker and many public/private clouds, and other infrastructure platforms supported by the community. You can do things like create multiple storage volumes for each user, or provision additional services on other containers/VMs.
Python 3.6 or above and JupyterHub 1.0.0 or above are required.
Example jupyterhub_config.py spawner configuration.
ansible_path = "/path/to/"
c.JupyterHub.spawner_class = "ansible"
c.AnsibleSpawner.inventory = ansible_path + "inventory.yml.j2"
c.AnsibleSpawner.create_playbook = ansible_path + "create.yml"
c.AnsibleSpawner.update_playbook = ansible_path + "update.yml"
c.AnsibleSpawner.poll_playbook = ansible_path + "poll.yml"
c.AnsibleSpawner.destroy_playbook = ansible_path + "destroy.yml"
c.AnsibleSpawner.playbook_vars = {
"container_image": "docker.io/jupyter/base-notebook",
"ansible_python_interpreter": "python3",
}
c.AnsibleSpawner.start_timeout = 600
c.JupyterHub.hub_connect_ip = "10.0.0.1"
Example playbooks and configurations can be found under examples.
Pytest is used to run automated tests that require Docker and Podman. These platforms were chosen because they are self-contained and can be installed in Travis, whereas testing with public cloud platforms requires secure access credentials.
If you only have one of these you can limit tests by specifying a marker. For example, to disable the Docker tests:
pytest -vs -m "not docker"
To view test coverage run pytest with --cov=ansiblespawner --cov-report=html, then open htmlcov/index.html.
setuptools-scm is used to manage versions. Just create a git tag.